| graph-classification-on-cifar10-100k | ESA (Edge set attention, no positional encodings) | #6 | Accuracy (%): 75.413±0.248 |
| graph-classification-on-dd | ESA (Edge set attention, no positional encodings) | #3 | Accuracy: 83.529±1.743 |
| graph-classification-on-enzymes | ESA (Edge set attention, no positional encodings) | #1 | Accuracy: 79.423±1.658 |
| graph-classification-on-imdb-b | ESA (Edge set attention, no positional encodings) | #2 | Accuracy: 86.250±0.957 |
| graph-classification-on-malnet-tiny | ESA (Edge set attention, no positional encodings) | #1 | Accuracy: 94.800±0.424MCC: 0.935±0.005 |
| graph-classification-on-mnist | ESA (Edge set attention, no positional encodings, tuned) | #1 | Accuracy: 98.917±0.020 |
| graph-classification-on-mnist | ESA (Edge set attention, no positional encodings) | #3 | Accuracy: 98.753±0.041 |
| graph-classification-on-nci1 | ESA (Edge set attention, no positional encodings) | #2 | Accuracy: 87.835±0.644 |
| graph-classification-on-nci109 | ESA (Edge set attention, no positional encodings) | #2 | Accuracy: 84.976±0.551 |
| graph-classification-on-peptides-func | ESA + RWSE (Edge set attention, Random Walk Structural Encoding, + validation set) | #1 | AP: 0.7479 |
| graph-classification-on-peptides-func | ESA + RWSE (Edge set attention, Random Walk Structural Encoding, tuned) | #3 | AP: 0.7357±0.0036 |
| graph-classification-on-peptides-func | ESA (Edge set attention, no positional encodings, tuned) | #13 | AP: 0.7071±0.0015 |
| graph-classification-on-peptides-func | ESA (Edge set attention, no positional encodings, not tuned) | #19 | AP: 0.6863±0.0044 |
| graph-classification-on-proteins | ESA (Edge set attention, no positional encodings) | #4 | Accuracy: 82.679±0.799 |
| graph-regression-on-esr2 | ESA (Edge set attention, no positional encodings) | #1 | R2: 0.697±0.000RMSE: 0.486±0.697 |
| graph-regression-on-f2 | ESA (Edge set attention, no positional encodings) | #2 | R2: 0.891±0.000RMSE: 0.335±0.891 |
| graph-regression-on-kit | ESA (Edge set attention, no positional encodings) | #2 | R2: 0.841±0.000RMSE: 0.433±0.841 |
| graph-regression-on-lipophilicity | ESA (Edge set attention, no positional encodings) | #6 | RMSE: 0.552±0.012R2: 0.809±0.008 |
| graph-regression-on-parp1 | ESA (Edge set attention, no positional encodings) | #1 | R2: 0.925±0.000RMSE: 0.343±0.925 |
| graph-regression-on-pcqm4mv2-lsc | ESA (Edge set attention, no positional encodings) | #1 | Validation MAE: 0.0235Test MAE: N/A |
| graph-regression-on-peptides-struct | ESA + RWSE (Edge set attention, Random Walk Structural Encoding, tuned) | #1 | MAE: 0.2393±0.0004 |
| graph-regression-on-peptides-struct | ESA (Edge set attention, no positional encodings, not tuned) | #8 | MAE: 0.2453±0.0003 |
| graph-regression-on-pgr | ESA (Edge set attention, no positional encodings) | #1 | R2: 0.725±0.000RMSE: 0.507±0.725 |
| graph-regression-on-zinc | ESA + rings + NodeRWSE + EdgeRWSE | #1 | MAE: 0.051 |
| graph-regression-on-zinc-500k | ESA + rings + NodeRWSE + EdgeRWSE | #1 | MAE: 0.051 |
| graph-regression-on-zinc-full | ESA + rings + NodeRWSE + EdgeRWSE | #1 | Test MAE: 0.0109±0.0002 |
| graph-regression-on-zinc-full | ESA + RWSE + CY2C (Edge set attention, Random Walk Structural Encoding, clique adjacency, tuned) | #2 | Test MAE: 0.0122±0.0004 |
| graph-regression-on-zinc-full | ESA + RWSE (Edge set attention, Random Walk Structural Encoding, tuned) | #4 | Test MAE: 0.0154±0.0001 |
| graph-regression-on-zinc-full | ESA + RWSE (Edge set attention, Random Walk Structural Encoding) | #5 | Test MAE: 0.017±0.001 |
| graph-regression-on-zinc-full | ESA (Edge set attention, no positional encodings) | #9 | Test MAE: 0.027±0.001 |
| molecular-property-prediction-on-esol | ESA (Edge set attention, no positional encodings) | #1 | RMSE: 0.485±0.009R2: 0.944±0.002 |
| molecular-property-prediction-on-freesolv | ESA (Edge set attention, no positional encodings) | #1 | RMSE: 0.595±0.013R2: 0.977±0.001 |